I think it would be best to stick with the spec for “retry-after” header. It appears that there are some custom headers that are being passed from the response. But a lot of http libraries already utilize this retry after header, and would make it more effective to implement openai requests.
Thank you,
Nate
MDN has a good doc on it.